The overall goal of this research is to advance our knowledge of the electrophysiological substrate and mechanism(s) of atrial fibrillation (AF) as evaluated at multiple endocardial recording sites. Statistical signal processing will be used to understand how the organizational structure of AF changes over time, with various interventions, and just prior to spontaneous termination. We posit that interventions which improve the degree of spatiotemporal organization of AF will reduce the endocardial atrial defibrillation theshhold (ADFT) and make AF more likely to spontaneously terminate, results which may have important implications for non-pharmacologic strategies of AF therapy. We will create two closed chest dog models AF by atrial enlargement and by rapid atrial pacing in order to show that shortening in total atrial effective refractory period (ERP) and increased spatial dispersion thereof leads to an increased propensity to sustain AF. We will show that vulnerability to induction of AF is related to regional dispersion of ERP. To show that regional and global spatiotemporal organization of AF are greater before spontaneous termination and successful defibrillation shocks, we will implement novel methods to measure parameters termed "persistence," and "instantaneous disorganization" based on the spatial extent and period of time over which activation sequence remains nearly constant. We will also test the hypothesis that atrial pacing resulting in regional entrainment during AF, by improving the degree of spatiotemporal organization, significantly reduces the ADFT. Finally, we will create a series of long linear atrial lesions guided by intracardiac echocardiography and test the hypothesis that catheter-based long linear RF lesions placed in the right atrium will result in a)a more organized pattern of atrial fibrillation; and b) a lower ADFT.